Background: Prostate cancer is a common disease that predominantly affects older men. Men with low-risk disease are frequently good candidates for observation, while those with high-risk disease and otherwise favorable life expectancy are often good candidates for treatment. In both situations, excellent cancer control with low morbidity can be achieved by appropriately selecting patients for observation versus treatment. Yet the management of prostate cancer seldom meets this ideal. High rates of both over and undertreatment exist. Hypotheses posited to explain these observations include a fragmented healthcare delivery system and lack of skill in prognostication among specialists involved in prostate cancer treatment decisions. In contrast, geriatric models of care seek to integrate care across providers and settings, promote shared decision-making, and use models of functional-status informed prognostication. Medicare claims-based analyses of prostate cancer treatment patterns offer the only nationally-representative data on the causes and consequences of over and undertreatment. Critical barriers to a deeper understanding of these phenomena using claims are lack of information regarding the provider networks responsible for beneficiaries' care and lack of clinically relevant information about beneficiary functional status. Objective: The overarching goal of this proposal is to address these critical barriers by: 1) developing empiric measurements of provider accountability and care attribution; and 2) exploring the influence of functional status and health-related quality of life on treatment decision making. Specific aims: 1) To empirically derive physician-hospital networks (PHNs) within SEER-Medicare data for future characterization of prostate cancer treatment patterns and outcomes among the elderly; and 2) to explore the effect of functional status on prostate cancer treatment patterns among the elderly. Study Design: We will use SEER-Medicare to develop primary care and prostate cancer treatment anchored PHNs and then profile PHN attributes hypothesized to be relevant to rates of over and undertreatment (Aim 1). We will use SEER-Medicare Health Outcomes Survey data and incorporate generalized estimating equations to examine patient-reported, geriatric-focused measures before and after prostate cancer treatment (Aim 2). Relevance to National Institute on Aging: Findings from this proposal will address critical barriers to claims- based analyses related to provider accountability/care attribution and to the lack of information on patient- reported measures. The proposal will provide geriatric-focused mentorship and methodological experience, including exposure to PHNs and to functional status measurement. These opportunities will lay the foundation for examining provider factors associated with better treatment targeting and for evaluating the relationship between functional status and prostate cancer treatment decision making.